4  Appendix A

4.1 Setup

4.1.1 Install Packages

We install the following packages using the groundhog package manager to increase computational reproducibility.

options(repos = c(CRAN = "https://cran.r-project.org")) 

if (!requireNamespace("groundhog", quietly = TRUE)) {
  install.packages("groundhog")
}

pkgs <- c("magrittr", "data.table", "stringr", "Rmisc", "gt")

groundhog::groundhog.library(pkg = pkgs,
                             date = "2024-08-01")

rm(pkgs)

4.1.2 Read Data

data      <- readRDS(file="../data/processed/full.Rda")

vars <- str_subset(string = names(data), pattern = "^.{1,4}$", negate = TRUE)
covariates <- data[stage == 1, ..vars]

4.2 Table A.1

long_df <- melt(covariates, 
                id.vars = c("surprise", "communication"), 
                measure.vars = c("age_18_34", "age_35_52", "age_53_plus", "female", 
                                 "high_education", "high_income", "married", "parentship", 
                                 "high_temperature", "high_usage", "high_general_risk", 
                                 "high_weather_risk", "high_accuracy", "high_credibility",
                                 "temperature", "usage", "general_risk", 
                                 "weather_risk", "accuracy", "credibility"),
                variable.name = "Variable", 
                value.name = "Value")

pooled_summary <- long_df[, .(N = sum(!is.na(Value)),
                              Mean = mean(Value, na.rm = TRUE)), 
                          by = .(Variable)]

summary_tmp <- long_df[, .(N = sum(!is.na(Value)),
                             Mean = mean(Value, na.rm = TRUE)), 
                         by = .(surprise, communication, Variable)]

summary_table <- merge(pooled_summary, summary_tmp, by = "Variable", suffixes = c("_pooled", ""))


summary_table_wide <- dcast(summary_table, Variable + N_pooled + Mean_pooled ~ surprise + communication, 
                            value.var = c("N", "Mean"))

setcolorder(summary_table_wide, 
            c("Variable", "N_pooled", "Mean_pooled", 
              "N_FALSE_point", "Mean_FALSE_point", 
              "N_FALSE_interval", "Mean_FALSE_interval", 
              "N_FALSE_both", "Mean_FALSE_both", 
              "N_TRUE_point", "Mean_TRUE_point", 
              "N_TRUE_interval", "Mean_TRUE_interval", 
              "N_TRUE_both", "Mean_TRUE_both"))

summary_table_wide %>%
  gt() %>%
  cols_label(
    Variable = "Treatment Variable",
    N_pooled = "N", Mean_pooled = "Mean",
    N_TRUE_point = "N", Mean_TRUE_point = "Mean",
    N_TRUE_interval = "N", Mean_TRUE_interval = "Mean",
    N_TRUE_both = "N", Mean_TRUE_both = "Mean",
    N_FALSE_point = "N", Mean_FALSE_point = "Mean",
    N_FALSE_interval = "N", Mean_FALSE_interval = "Mean",
    N_FALSE_both = "N", Mean_FALSE_both = "Mean") %>%
  tab_spanner(
    label = "Pooled",
    columns = c(N_pooled, Mean_pooled)) %>%
  tab_spanner(
    label = "Point",
    columns = c(N_TRUE_point, Mean_TRUE_point, N_FALSE_point, Mean_FALSE_point)) %>%
  tab_spanner(
    label = "Interval",
    columns = c(N_TRUE_interval, Mean_TRUE_interval, N_FALSE_interval, Mean_FALSE_interval)) %>%
  tab_spanner(
    label = "Point + Interval",
    columns = c(N_TRUE_both, Mean_TRUE_both, N_FALSE_both, Mean_FALSE_both)) %>%
  tab_spanner(
    label = "Confirmation",
    columns = c(N_FALSE_point, Mean_FALSE_point, N_FALSE_interval, Mean_FALSE_interval, N_FALSE_both, Mean_FALSE_both)) %>%
  tab_spanner(
    label = "Contradiction",
    columns = c(N_TRUE_point, Mean_TRUE_point, N_TRUE_interval, Mean_TRUE_interval, N_TRUE_both, Mean_TRUE_both)) %>%
  fmt_number(
    columns = starts_with("Mean"),
    decimals = 3) %>%
  fmt_number(
    columns = starts_with("N"),
    decimals = 0) %>%
  cols_align(
    align = "left",
    columns = c(Variable)
  )

Table 4.1: Descriptive statistics (control variables): Mean values per treatment

			Contradiction		Confirmation		Contradiction		Confirmation		Contradiction		Confirmation
Treatment Variable	Pooled		Point				Interval				Point + Interval
	N	Mean	N	Mean	N	Mean	N	Mean	N	Mean	N	Mean	N	Mean
age_18_34	1,503	0.335	251	0.386	255	0.325	258	0.345	243	0.309	249	0.345	247	0.296
age_35_52	1,503	0.306	251	0.299	255	0.310	258	0.302	243	0.313	249	0.313	247	0.300
age_53_plus	1,503	0.359	251	0.315	255	0.365	258	0.353	243	0.379	249	0.341	247	0.405
female	1,505	0.490	252	0.540	255	0.463	258	0.484	243	0.444	250	0.528	247	0.478
high_education	1,505	0.618	252	0.643	255	0.616	258	0.578	243	0.687	250	0.612	247	0.575
high_income	1,363	0.525	225	0.547	230	0.552	234	0.517	220	0.564	227	0.471	227	0.498
married	1,505	0.396	252	0.389	255	0.427	258	0.403	243	0.379	250	0.356	247	0.421
parentship	1,505	0.429	252	0.440	255	0.400	258	0.450	243	0.395	250	0.404	247	0.482
high_temperature	1,505	0.591	252	0.575	255	0.557	258	0.589	243	0.613	250	0.668	247	0.543
high_usage	1,505	0.511	252	0.548	255	0.518	258	0.465	243	0.494	250	0.516	247	0.526
high_general_risk	1,505	0.540	252	0.512	255	0.561	258	0.589	243	0.506	250	0.552	247	0.514
high_weather_risk	1,505	0.633	252	0.679	255	0.600	258	0.678	243	0.588	250	0.660	247	0.587
high_accuracy	1,505	0.786	252	0.698	255	0.800	258	0.756	243	0.819	250	0.836	247	0.810
high_credibility	1,505	0.503	252	0.369	255	0.490	258	0.484	243	0.560	250	0.504	247	0.615
temperature	1,505	20.051	252	20.079	255	19.906	258	20.062	243	20.267	250	20.468	247	19.526
usage	1,505	3.347	252	3.393	255	3.408	258	3.275	243	3.350	250	3.308	247	3.348
general_risk	1,505	4.193	252	4.048	255	4.290	258	4.407	243	4.012	250	4.396	247	3.988
weather_risk	1,505	5.260	252	5.433	255	5.122	258	5.585	243	4.967	250	5.496	247	4.935
accuracy	1,505	2.241	252	1.992	255	2.306	258	2.081	243	2.366	250	2.308	247	2.405
credibility	1,505	2.378	252	2.056	255	2.380	258	2.380	243	2.498	250	2.404	247	2.559